Websites have become increasingly complex venues for commerce (e.g., the buying and selling of goods and services over the Internet) and now include interactive functionality and sophistication that allows communication between users and a variety of databases. As such, it has become increasingly difficult to manage transitions from one software revision of a website to another software revision of the website without the use of engineers that test quality and stability of the website prior to release. Often times, testing a portion of a website affects operations within a remainder of the website because changing parameters and variables of the portion may affect the input variables of the remainder (e.g., new code can sometime break old functionality because files being changed have input and output parameters that affect the stability of the remainder). Furthermore, multiple engineers may work on the same portion of software code and a particular engineer may inadvertently modify a parameter that will affect the stability of the remainder (e.g., an engineering team may be focused on solving one problem, and may inadvertently create new problems by changing parameters that are required by another portion of software code associated with the website).
Technology to aid testing of software revisions of websites has largely been limited to external testing tools that require an engineering team to learn a special programming language to define what aspects of the website will be tested (e.g., SILKCENTRAL™ Test Manager by Segue Software, Inc., and MERCURY WINRUNNER™ by Mercury Interactive, Inc.), and internally developed testing systems that are dependent upon manually testing all aspects of a software revision prior to release. External testing tools require that a website owner hire specialized engineers that are trained in specialized programming languages associated with the external tools. Furthermore, external tools cannot test aspects (e.g., physical locations of images, text, or input boxes on a web page) of a software revision that have not been hard coded as candidates for testing (e.g., an engineer must predetermine which aspects of a website are to be tested and hard code instructions for testing the predetermined aspects). Manual offline testing systems are error prone because engineers may not be able to test all aspects of a software revision prior to release (e.g., engineers may make a mistake, or may not have the manpower needed to test every aspect of a website).
In order to make software testing systems more efficient, there is some incentive for operators to provide systems for automating the testing of software revisions associated with websites without the use of external testing tools that require hard coding of testing candidates, and without hiring specialized engineers that are trained in specialized programming languages associated with the external tools. However, the design of such software testing systems present a number of technical challenges, specifically relating to which data is examined as a testing candidate, how to reduce the number of false positives if HTML data is used as the testing candidate, how accuracy is maintained, and how hardware architecture is designed and/or implemented to manage automated software testing systems.